BioHackathon 2009

Towards integrated web service in Open Bio*

Objectives

Dear all,

Database Center for Life Science (DBCLS) is planning to have a BioHackathon? for the Web Service in next January or February and hope to invite all of you to Japan (currently considering upto 20 developers from abroad and other 20 developers from Japan in total). We can offer your travel fare and accommodation fee by the DBCLS budget.

DBCLS ( http://dbcls.jp/en/) is a non-profit research institution funded by MEXT ( http://www.mext.go.jp/english/), established on the April 2007 and will continue for the 4 years. One of our goals is to integrate bioinformatics web services in a unified and convenient manner.

For this purpose, I'm planning to develop a proxy server which translate user's query for appropriate web service provider (e.g. EBI, KEGG, BioMOBY etc.) and formalize the returned data structures, so that user can easily build their own workflow without bothered by the data format conversion.

This view is partly based on the comment by Tom Oinn in BOSC2007 - even a sequence object in BioPerl? is not compatible with other Bio* libraries - I strongly agree with it. We should define "50 basic bioinformatics data types" and implement them by refactoring current Bio* libraries, then use them in every web services.

BioMOBY has been served as a initial repository for various web services, however, there are hundreds of data types are listed and many of them are very similar. I think they need to be re-organized as the naming convention of the methods and data types currently used are extremely diverged. I suppose this situation restrict the interoperability of the services, and to develop any bioinformatics workflow, user needs to consult with (poor, in some cases) documentations.

So, I hope to gather as many developers as we can, from major web service providers, from Bio* developers, and other integrated service providers like BioMOBY and Taverna, to discuss and develop a new Open Bio* standard for the web-service-centric bioinformatics-nation (TM).

I'm preferentially writing you to hear any input to finalize our objectives of the web service hackathon, as my current knowledge on BioMOBY, myGrid, Taverna and other web services are very limited.

From my experience as a developer of the BioRuby library and the KEGG API service, I'm still stick on the SOAP/WSDL services even though the REST is the current trend. Because,

• every SOAP server speaks the same language
  • cost to develop a client code can be minimal
  • user doesn't need any parser (cf. DAS spec etc.)
  • user can easily pass any complex data structure (including <ComplexType?>)
• we can also provide REST service if appropriate

On behalf of the DBCLS, Toshiaki Katayama

Venue

• Tokyo, Japan
• Candidate dates

1. 2008/02/10-02/16
2. 2008/01/27-02/02
3. 2008/02/03-02/09

Attendees

• Toshiaki Katayama (Univ of Tokyo, Japan; BioRuby, KEGG API, KEGG DAS)
• Alberto Labarga (Integromics, Spain; ex-EBI)
• Matthew Pocock (Newcastle, UK; Taverna, BioJava?)
• Sophia Ananiadou (Manchester, UK; Text mining)

If you could recommend any appropriate persons from the following groups, please let me know:

• NCBI web service developers
• BioMOBY service developers (I found some interesting providers in Spain)
• BioPerl?, BioPython?, and BioJava? developers

Starting Points

• TracGuide -- Built-in Documentation
•  The Trac project -- Trac Open Source Project
•  Trac FAQ -- Frequently Asked Questions
• TracSupport -- Trac Support

For a complete list of local wiki pages, see TitleIndex.